In this fiscal year, we continued our investigation on the folding mechanism of barnase, a ribonuclease, which has been a classical model for protein folding. In contrast to the generally accepted conclusion that barnase folds through an early folding intermediate, we found convincing experimental evidences that show absence of stable intermediate. Our finding is important in several aspects (i) It further validates the native-state HX approach (Bai et al., Science, 1995). (ii) It addresses the importance of using multiple approaches to characterize folding pathways. (iii) It supports the view proposed recently by us and others that small proteins (<~ 100 AA) fold by a nucleation mechanism without population of early folding We also continued our test on the method of using phage-display and proteolysis to select folded stable proteins. We are able to select stable four-helix bundle proteins based on a protein library constructed from a partially unfolded protein, apocytochrome b562. Thermodynamic studies of the selected proteins suggest that this method is efficient to select stable proteins. Spectroscopic studies suggest that the selected proteins also fold into a four-helix conformation. An atomic resolution structure of the most stable protein selected will be determined in the future by NMR to see whether the protein has an unique hydrophobic core. These results demonstrate that the method tested here should be very useful to engineer stable proteins and to design new protein. Taking advantage of the success of the phage-display approach to select stable proteins, we have studied the folding pathway of the stable four-helix bundle proteins. We found that the four-helix bundle proteins fold without population of an early folding intermediate. Even importantly, we found that an intermediate exist after the rate-limiting transition state.